Hi,
We are using heartbeat 2.1.3 from the SLES 10 SP2 x64 distribution and are
trying to manage 2 drbd resource agents in a 2 node cluster. Drbd is version
8.3.2 and we are using the updated ocf resource script that is bundled with
DRBD 8.3.2 which we renamed to drbd832.
We configured the first drbd resource in the master_slave block and it gets
promoted to master fine, yet whenever we add a second drbd resource to the same
master_slave block neither resource sets a node as master. Both DRBD resources
are in a group and the group resides within the master_slave section of the
CIB. Both drbd resources start on both nodes, but neither node is promoted to
master . We also noticed that master-slave meta attributes do not seem to
apply to our group members. We had to add individual meta_attributes to each
primitive.
Using drbdadm on the command line, we are able to promote/demote either node to
primary/secondary with no problems.
We would like to know if configuring more than one master-slave resources works
in 2.1.3 and how should this be configured. Do we need individual master-slave
blocks in the CIB?
Here is an excerpt from our CIB:
<master_slave id="ms_drbd">
<meta_attributes id="ms_drbd-ma">
<attributes>
<nvpair id="ma-ms-drbd0-1" name="clone_max" value="2" />
<nvpair id="ma-ms-drbd0-2" name="clone_node_max" value="1" />
<nvpair id="ma-ms-drbd0-3" name="master_max" value="1" />
<nvpair id="ma-ms-drbd0-4" name="master_node_max" value="1" />
<nvpair id="ma-ms-drbd0-5" name="notify_start_uname" value="qpr1" />
<nvpair id="ma-ms-drbd0-6" name="notify" value="yes" />
<nvpair id="ma-ms-drbd0-7" name="globally_unique" value="false" />
<nvpair id="ma-ms-drbd0-8" name="target_role" value="#default" />
</attributes>
</meta_attributes>
<group id="drbd-group">
<primitive id="drbd-var-mqm" class="ocf" provider="heartbeat" type="drbd832">
<meta_attributes id="ms_drbd1-ma">
<attributes>
<nvpair id="ma-ms-drbd1-1" name="clone_max" value="2" />
<nvpair id="ma-ms-drbd1-2" name="clone_node_max" value="1" />
<nvpair id="ma-ms-drbd1-3" name="master_max" value="1" />
<nvpair id="ma-ms-drbd1-4" name="master_node_max" value="1" />
</attributes>
</meta_attributes>
<instance_attributes id="ms_drbd-var-mqm-instance_attributes">
<attributes>
<nvpair id="drbd-b-1" name="drbd_resource" value="var-mqm" />
<nvpair id="drbd-b-2" name="target_role" value="started" />
<nvpair id="drbd-b-3" name="is_managed" value="true" />
</attributes>
</instance_attributes>
</primitive>
<primitive id="drbd-var-mqm-log" class="ocf" provider="heartbeat"
type="drbd832">
<meta_attributes id="ms_drbd2-ma">
<attributes>
<nvpair id="ma-ms-drbd2-1" name="clone_max" value="2" />
<nvpair id="ma-ms-drbd2-2" name="clone_node_max" value="1" />
<nvpair id="ma-ms-drbd2-3" name="master_max" value="1" />
<nvpair id="ma-ms-drbd2-4" name="master_node_max" value="1" />
</attributes>
</meta_attributes>
<instance_attributes id="ms_drbd-instance_attributes">
<attributes>
<nvpair id="drbd-a-1" name="drbd_resource" value="var-mqm-log" />
<nvpair id="ms_drbd-target_role" name="target_role" value="started" />
<nvpair id="ms_drbd-is_managed" name="is_managed" value="true" />
</attributes>
</instance_attributes>
</primitive>
</group>
</master_slave>
crm_mon output:
============
Last updated: Mon Oct 12 11:25:27 2009
Current DC: qpr2 (e74cb8c7-50a7-41d4-80af-abd4da061418)
2 Nodes configured.
2 Resources configured.
============
Node: qpr2 (e74cb8c7-50a7-41d4-80af-abd4da061418): online
Node: qpr1 (08ee8c47-dac6-4fa4-b157-0dcf4901d469): online
Master/Slave Set: ms_drbd
Resource Group: drbd-group:0
drbd-var-mqm:0 (ocf::heartbeat:drbd832): Started qpr2
drbd-var-mqm-log:0 (ocf::heartbeat:drbd832): Started qpr2
Resource Group: drbd-group:1
drbd-var-mqm:1 (ocf::heartbeat:drbd832): Started qpr1
drbd-var-mqm-log:1 (ocf::heartbeat:drbd832): Started qpr1
Resource Group: xpr-group
app_addr (ocf::heartbeat:IPaddr2): Started qpr2
public_addr (ocf::heartbeat:IPaddr2): Started qpr2
default_route (ocf::heartbeat:Route): Started qpr2
Thanks for taking a look. We've been spinning our wheels on this one.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
